Method and Materials
Pollen identification, counting and definition of pollen season
Pollens were identified and counted from Hanyang University Seoul and Guri hospitals for over 22 years (1998~2019). Pollen sampling has been undertaken at the sites with Burkard Seven-days Sampler (Burkard Manufacturing Co Ltd., Hertfordshire, UK) on exposed roof tops during the study. Cumulative pollen concentrations and a preset percentage threshold level are used to define the start and end of the pollen season. The present study applied the 5 % threshold method, which gave the smallest standard deviation in the start dates.13
Selection of subject and sensitization rate to allergenic tree pollen
Subjects for this study were selected by using the allergy skin prick test and the quantitative serum test for allergen-specific IgE level. Over 3 class levels of a pollen allergen-specific serum IgE or > 3+ level of allergic skin prick test to pollen are an indication of sensitization to the allergenic pollen.14 Pollen for allergy test were birch, alder, oak, hazel, pine, willow. Sensitization rate to allergenic pollen was defined as a percentage of the number of the patients who sensitized allergenic pollen in the overall population taken allergy tests. Five age group were divided for the evaluation of change of age distribution (3~5, 6~9, 10~12, 13~15, and 16~18 age of years).
Meteorological data
Daily records of the climatic variables such as mean temperature, rainfall, accumulated temperature and 7-days sunshine amount from the Korea Meteorological Administration (KMA) were used to represent the daily weather conditions near to the pollen monitoring site. The nearest weather stations to each pollen sampling site were selected to provide a number of meteorological data. The accumulated temperature defined the summed up temperature from 1st January until the starting day of pollination, which is the most influential variables to pollination among meteorological variables and was selected for this study.
Statistical analysis
We aimed to assess long-term pollen data to ascertain potential links to sensitization rate during the collection period. R version 4.0.2 (The R foundation, https://www.r-project.org) was used to ascertain statistical analysis. The changes of pollen sensitization were assessed by linear regression and time series analysis. A standardized regression coefficient by a method of least square for minimizing the sum of squared residuals is estimated, and which independent variable has a large influence on the dependent variable is found out.15
The correlation between the duration of the pollen season with starting date and pollen sensitization was assessed by correlation analysis. Long-term prediction of variation of pollen season was assessed by ARIMA-11 analysis. Time series data are provided after seasonal adjustment. The ARIMA-11 analysis may also backcast previous year data from the starting point of the data, predict the year data from the ending point of the data, and solve the missing value problem by moving average of the expanded data.16
Ethics statement
This study was approved by the institutional review board of the Hanyang University Guri Hospital (IRB No. HYI-10-45). The requirement for informed consent from patients was waived, as the study was a retrospective analysis.